Urbanization, renewable energy production, and carbon dioxide emission in BSEC member states: implications for climate change mitigation and energy markets

Abstract

As the world’s population grows, the energy demand continues to rise due to advancements in technology and the impact of globalization. The finite nature of traditional energy sources has accelerated the shift toward renewable energy, particularly in developing countries where environmental degradation and declining quality of life are significant concerns. This study delves into the interplay between urbanization, carbon dioxide emissions, economic growth, and renewable energy production in Organization of the Black Sea Economic Cooperation member states, providing new insights into the energy market. By using annual data from 1995 to 2020 and advanced panel cointegration tests, this study provides a comprehensive analysis of the determinants of renewable energy for developing countries. The findings show a substantial and long-term relationship between urbanization, emissions, growth, and renewable energy production. These findings have important implications for policymakers and underscore the critical role of renewable energy in mitigating climate change in developing countries.

Introduction

The ongoing global energy crisis is significantly influenced by the problem of climate change and energy markets. Even if things may appear hopeless, a green transition has gathered steam in recent years due to an increase in investments in renewable energy that offers some ray of hope. But because coal is the most polluting of all the fossil fuels (Khan, Han et al. 2021) and natural gas costs have increased, there is an urgent need to improve energy efficiency and offer more robust government incentives to encourage the expansion of renewable energy.

On the plus side, for the first time last year, global capital expenditure on wind and solar energy assets outpaced investment in new and existing oil and gas wells. The green transition is being sped up by five to 10 years because industrialized countries’ governments offer billions in subsidies to encourage investments in clean technologies. To restrict temperature increases to 1.5 °C over pre-industrial levels by 2100, it is essential to achieve net-zero carbon emissions by 2050. However, more work needs to be done in this area. By 2030, global renewable energy investment will need to triple to almost $4 trillion to achieve net zero emissions by 2050 (https://www.iea.org/reports/net-zero-by-2050).

It has been difficult to create a coherent and integrated solution for energy policies due to the intricate and fragmented nature of the energy markets in the BSEC economies. Despite the region’s large deposits of oil, gas, and coal, energy security is still a top priority for many of the region’s nations. Due to their reliance on energy imports from outside the area, some nations are vulnerable to supply disruptions and price fluctuations.

Significant differences in the price and accessibility of energy across the BSEC region are partly a result of the fragmented structure of the energy markets. Regulatory obstacles and weak infrastructure make it difficult for countries with ample energy resources to transport and export their resources. Nevertheless, countries with a dearth of indigenous resources are compelled to rely on expensive energy imports, which can negatively influence their economy and energy security. Renewable energy development has emerged as a potential solution to address some of the energy challenges in the region. The BSEC countries have significant potential for renewable energy, including solar, wind, and hydropower. However, the development of renewable energy has been limited by various factors, including a lack of investment and inadequate regulatory frameworks.

Despite progress, the renewable sector faces new challenges, such as price caps, taxes, and rising costs, rendering energy projects less appealing. Additionally, logistical disruptions and higher interest rates increase the cost of developing green plants. These challenges must be addressed to ensure that renewable energy remains a promising investment and that the world can realize a more sustainable energy future.

Climate change is also a pressing issue that cannot be ignored, and the need to transition toward cleaner energy sources is crucial in mitigating its impacts. Throughout human history, there has been a continuous need for energy due to various reasons, leading to a significant increase in energy demand.

Individual energy needs have particularly grown during globalization, especially with the removal of restrictions on capital following the Industrial Revolution (Shahbaz et al. 2016). Significant changes have been made in utilizing energy resources to meet this growing demand. The supply and demand for goods and services have continuously increased from the beginning of the industrialization era to the present day. With the increasing population, the energy needs have also grown to a considerable extent. In the growth and development strategies of countries, the need for energy is significant, particularly for developing countries where a large majority is dependent on foreign energy. Therefore, energy efficiency, diversity, and preference for renewable energy over non-renewable energy sources have become crucial. As energy dependence on foreign sources increases, countries’ growth and development strategies can be hindered, and various limitations can be faced as a result.

Looking at the world’s population that is approaching 8 billion and is projected to reach approximately 9 billion by 2040, and taking into account technological advancements, it is clear that there will not be a decrease in energy demand but rather significant increases. However, the decreasing reserves of fossil fuels, the primary energy source referred to as non-renewable energy, pose significant problems in terms of meeting energy demand sustainably. With coal reserves projected to be depleted in approximately 150 years, natural gas reserves in approximately 60 years, and oil reserves in approximately 40 years, countries are turning toward alternative energy sources to meet their energy needs (Haider 2020; Zou et al. 2016).

Due to the constraints in traditional energy sources, the importance of renewable energy sources, which are sources that can renew themselves with certain processes, is increasing. Examples of renewable energy sources include solar, biomass, wind, hydro, and geothermal (Rahman et al. 2022). These sources are thought to play an essential role in solving specific negative impacts threatening present and future generations (Zhiznin et al. 2019). Renewable energy sources have been acknowledged as a crucial aspect in addressing various environmental and economic issues. With the increasing use of these sources, it is believed that the negative impacts on the environment, such as erosion and loss of biodiversity, can be reduced, as well as the damage caused to agricultural production and the global warming caused by emissions from traditional energy sources. It has become crucial for policymakers to identify the different types of renewable energy sources and understand the factors influencing their demand. The shift toward renewable energy not only brings environmental benefits such as a reduction in pollution but also contributes to the economic development of countries.

Additionally, it is vital to increase the efficiency of the existing energy supply. Governments are essential in implementing measures that increase energy efficiency and ensure efficient energy use. This is particularly important in developing countries, where energy costs constitute a significant proportion of the overall costs. Increasing the shift toward renewable energy and promoting energy efficiency can help mitigate these costs and ensure the continuity of the energy supply. The continuity of energy supply is crucial to protect these countries from external shocks, prevent deterioration of macroeconomic indicators, avoid production disruptions, and minimize problems that may arise in the supply chain.

The study is aimed at testing whether there is a long-term relationship between urbanization, carbon dioxide emissions, economic growth, and renewable energy production in BSEC (Black Sea Economic Cooperation) countries.

The theoretical contribution of this study is to provide evidence of a long-term relationship between urbanization, carbon dioxide emissions, economic growth, and renewable energy production in BSEC member states. This study fills a gap in the literature by using annual data from 1995 to 2020 and employing the second-generation panel cointegration tests to analyze the determinants of renewable energy sources for developing countries. The findings of this study have implications for policymakers in developing countries and can inform decision-making on renewable energy policies. Additionally, the study contributes to understanding the interplay between economic growth, urbanization, CO₂ emissions, and renewable energy production in the context of BSEC countries.

The paper consists of the following sections: the “Introduction” section, which provides background information on the topic, introduces the research problem and objectives and outlines the structure of the paper. The “Literature review” section discusses previous studies and research related to the topic, identifies gaps in the current knowledge, and sets the stage for the present study. The “Methodology” section relates to the statistical analysis procedures used in the study, model specification, and data source. The “Findings and discussion of the results” section consists of the study’s findings, including tables, figures, and graphs, and provides a detailed explanation of the data analysis. Finally, in the “Conclusion and recommendations” section, the study ends with the conclusions, recommendations, and areas for future research.

Literature review

Renewable energy and CO₂

Large number of studies have been conducted on the relationship between renewable energy and carbon dioxide (CO₂) emissions such as Zhu et al. (2019) modeled using variable renewable energy sources, such as wind and solar generators, in coupled electricity and heating systems to reduce CO₂ emissions. Ben Jebli et al. (2019) explored the causal relationships between renewable energy consumption, tourism, economic growth, foreign direct investment, and CO₂ emissions for a panel of Central and South American countries from 1995 to 2010. Likewise, Bilan et al. (2019) investigated the impact of renewable energy sources, CO₂ emissions, macroeconomics, and political stability on gross domestic product (GDP). Ullah et al. (2020) used a simultaneous equation approach for their analysis, while Nathaniel and Iheonu, (2019) employed the Augmented Mean Group (AMG) estimation technique to examine the relationship between renewable and non-renewable energy consumption and CO₂ abatement in Africa from 1990 to 2014. Fatima et al. (2021) analyzed the role of renewable and nonrenewable energy concerning the rising carbon emissions in leading emitting countries. Ben Jebli et al. (2020) looked at the relationship between CO₂ emissions, economic growth, renewable energy consumption, industrial and service value added, while Khan et al. (2021a, b) estimated the short-term and long-term impacts of technological innovation, finance, and foreign direct investment on renewable energy, non-renewable energy, and CO₂ emissions in the “Belt and Road Initiative (BRI)” countries. Acheampong et al. (2021) investigated the inter-temporal causal relationship between institutions, renewable energy, carbon emissions and economic growth for sub-Saharan Africa countries. Abban et al. (2022) argue that the shift toward renewable sources of energy not only that is beneficial for the natural ecosystems but also improves economic growth in African oil-producing economies. Mehmood (2021) attempted to investigate the impacts of renewable energy on CO₂ emissions by including factors such as education, GDP, natural resources, and foreign direct investment for G11 economies. Kerem (2022) investigated the carbon footprint effect during electricity generation in Kahramanmaraş, Turkey, with abundant renewable energy potential. Abbas et al. (2022) looked at the influence of renewable energy development, market regulation, and environment-related innovation on CO₂ emissions in the BRICS countries and found that there are both symmetric (direct) relationships between these variables along with asymmetric (indirect) ones mediated by market regulations which have an impact on CO₂ emissions in those countries, suggesting that policymakers should focus more on developing renewable energies, introducing environmental taxes while also expanding green technological innovations for better control over carbon dioxide emission levels which can ultimately lead toward greater sustainability goals. Additionally, Dogan et al. (2023) points out that fiscal instruments need to be implemented carefully, as energy taxes tend to negatively impact on renewable energy. At the same time, recent research (Hashmi et al. 2022) proves that expansionary monetary policy is essential in promoting the renewable energy transition. Nevertheless, energy efficiency should be also taken into consideration along with the renewable sources as it also promotes the mitigation of carbon emissions (Jahanger et al. 2023). Energy efficiency is essential as the transition toward renewable sources of energy cannot be done instantly and the path toward fully green-fueled economy is complex supposing the usage of traditional sources.

Mehmood (2022a, b) used cross-sectional autoregressive distributed lag (CS-ARDL) approach to know long- and short-run coefficient values, while Ali et al. (2022) by using Dynamic ARDL simulation technique explored the association among renewable energy and non-renewable energy consumption, urban population, research and development expenditure, technological innovation, and carbon emission intensity in Russia and found strong evidence of long run correlations between these variables; renewable energy consumption had a positive influence on carbon emission intensity by 0.27%. Izadi et al. (2023) simulated a hybrid renewable energy system with hydrogen energy storage to cover the energy demand of sections and wards of a hospital that dealt with COVID-19 patients. Qin et al. (2022) used the time-varying parameter-stochastic volatility-vector auto-regression (TVP-SV-VAR) model to obtain the changing relations among US sustainable finance, renewable energy, and carbon dioxide emission. Mehmood (2022a) investigates the influences of banking sector development on CO₂ emissions (CO₂e) in the next eleven countries. Studies have also focused on the relationship between sustainable development goal (SDG) and CO₂ emissions. The Sustainable Development Goal (SDG) aims to provide a practical solution for achieving sustainable growth by focusing on affordable and sustainable energy (L. Wang et al. 2022), in order to understand the factors that drive the consumption of renewable energy and the relationship between GDP growth, technological innovation, gross fixed capital formation, CO₂ emissions, income inequality, and renewable energy consumption (REC). Apergis et al.’s (2023) study is the first study to examine the relationship between CO₂ emissions and the consumption of renewable and non-renewable energy in Uzbekistan over a period of 1985–2020. Khan et al. (2020) studied the nexus between renewable energy consumption, carbon emission, and financial development in the global panel of 192 countries using panel quantile regression. The authors found that all the variables are heterogeneous across quantiles. The effect of renewable energy on carbon dioxide is negative while financial development is positive. The effect of carbon dioxide is negative on renewable energy while financial development has a positive effect on renewable energy.

Renewable energy and urbanization

A literature review of recent studies on the topic of renewable energy and urbanization shows a growing interest in understanding the relationship between these two factors and their impact on the environment. Khan et al. (2018) used panel data from 1990 to 2015 to explore the association between greenhouse gases, financial development, forest area, improved sanitation, renewable energy, urbanization, and trade in 24 lower middle-income countries from Asia, Europe, Africa, and America. The study found a positive correlation between renewable energy and improved sanitation and a negative correlation between renewable energy and greenhouse gas emissions. Salim et al. (2019) aimed to investigate the effects of urbanization on pollutant emissions and energy intensity in selected Asian developing countries. They controlled for the effects of disaggregated (renewable and non-renewable) energy consumption, trade liberalization, and economic growth and found that urbanization had a positive effect on pollutant emissions and energy intensity. Wang et al. (2021) investigated the dynamic interdependence between CO₂ emissions, real gross domestic product (GDP), renewable and non-renewable energy generation, urbanization, and export quality for both the top ten renewable energy and top ten economic complexity index (ECI) countries. They used the generalized method of moment (GMM) estimation on a data set ranging from 1993 to 2018 and found a dynamic interdependence between CO₂ emissions, real gross domestic product (GDP), renewable and non-renewable energy generation, urbanization, and export quality.

Ulucak and Khan (2020) examined the relationship between real income, renewable energy, urbanization, natural resource rent, and ecological footprint in BRICS economies. They found that an increase in renewable energy and urbanization was associated with a decrease in ecological footprint. Nathaniel et al. (2020) investigated the effects of renewable energy, urbanization, financial development, and economic growth on the ecological footprint in MENA countries. They applied the Augmented Mean Group algorithm and found that renewable energy had a positive effect on the ecological footprint.

Akadiri et al. (2022) examined the impacts of globalization, real income, urbanization, and energy consumption on environmental degradation in Nigeria, using quarterly frequency time series data over a period 1971–2018 and found that due to the use of fossil fuel combustion increases the presence of CO emissions and other greenhouse gases in the atmosphere in several countries of the world, including Nigeria. Xia (2022) explored the relationship between green technology innovation (GI) and renewable energy investment (REI) in selected Chinese provinces from 2005 to 2019 and found positive association between green technology innovation (GI) and renewable energy investment (REI). Rahman et al. (2022) investigated the impact of trade openness and urbanization effect on renewable and non-renewable energy consumption in emerging economies for the period 1990–2018, using different approaches such as structural break test, autoregressive distributed lag (ARDL)-bounds testing, and Granger causality, and found that trade openness and urbanization effect on renewable and non-renewable energy consumption in emerging economies.

One common theme among the studies is the investigation of the relationship between renewable energy and urbanization, and their impact on environmental outcomes such as greenhouse gas emissions, energy intensity, ecological footprint, and pollution. These studies have also considered other factors that may affect these outcomes, such as economic growth, trade openness, financial development, and natural resource rent.

Overall, the literature suggests that increasing renewable energy consumption and promoting sustainable urbanization can have positive effects on the environment. However, further research is needed to better understand the specific mechanisms and policies that can effectively promote the adoption of renewable energy and sustainable urbanization in different contexts.

Renewable energy and economic growth

A literature review of the relationship between renewable energy (RE), non-renewable energy (NRE), and economic growth reveals that the majority of studies suggest that RE has a positive effect on economic growth, while NRE has a negative effect. Majority of the authors (Awodumi and Adewuyi 2020; Bekun et al. 2019; Kahia et al. 2016; Koçak and Şarkgüneşi, 2017; Luqman et al. 2019; Maji et al. 2019; Ntanos et al. 2018; Omri et al. 2015; Pao and Fu 2013; Zaidi et al. 2018) utilized the autoregressive distributed lag (ARDL) approach and found that RE had a positive effect on economic growth. Additionally, NRE had a negative effect on economic growth, as well as a positive effect on carbon emissions in some cases. These findings suggest that increasing the use of renewable energy could be beneficial for economic growth while reducing the use of non-renewable energy could be beneficial for both economic growth and carbon emissions.

Emir and Bekun (2019) studied the relationship between energy intensity, carbon emissions, renewable energy, and economic growth in Romania. The NAR-DLM was used to analyze the data, and the results showed that energy intensity and carbon emissions have a negative effect on economic growth and a positive effect on renewable energy consumption.

Aydoğan and Vardar (2020) studied the role of renewable energy, economic growth, and agriculture on CO₂ emission in E7 countries. The EKC model was used to analyze the data, and the results showed that renewable energy consumption has a positive effect on economic growth and a negative effect on carbon emissions. Mahmood et al. (2019) studied the relationship between renewable energy, economic growth, human capital, and CO₂ emission. The results showed that renewable energy consumption has a positive effect on economic growth and a negative effect on carbon emissions. Zafar et al. (2019) examined the relationship between renewable and nonrenewable energy consumption, trade openness, and CO₂ emissions in the framework of EKC. The results showed that renewable energy consumption has a positive effect on economic growth and a negative effect on carbon emissions.

Ridzuan et al. (2020) studied the effects of agriculture, renewable energy, and economic growth on carbon dioxide emissions. The EKC model was used to analyze the data, and the results showed that renewable energy consumption has a positive effect on economic growth and a negative effect on carbon emissions. Wang and Wang (2020) examined the relationship between renewable energy consumption and economic growth in OECD countries. The nonlinear panel data analysis was used to analyze the data, and the results showed that renewable energy consumption has a positive effect on economic growth. Abbas et al. (2020) studied the role of fixed capital formation, renewable and non-renewable energy in economic growth, and carbon emission. The results showed that renewable energy consumption has a positive effect on economic growth and a negative effect on carbon emissions.

While analyzing the 2022 articles relating to the topic, the articles analyzed use various quantitative methods such as composite indices, nonlinear ARDL, ARIMA-fGARCH, and MMQR estimations to study the effects of renewable energy on economic growth and environmental sustainability. Afroz and Muhibbullah (2022) used a nonlinear ARDL approach to investigate the dynamic linkages between non-renewable energy, renewable energy, and economic growth in Malaysia. Their findings revealed that non-renewable energy consumption had a positive effect on economic growth, while renewable energy consumption had a negative effect. Moreover, a policy mix that promotes both renewable sources of energy and technological innovation is desirable as it mitigates the environmental degradation (Suki et al. 2022). Fakher et al. (2022) used composite indices for environmental quality and financial development to study the relationship between renewable energy, financial development, environmental quality, and economic growth. Their results showed that renewable energy consumption had a positive effect on financial development and environmental quality, while non-renewable energy consumption had a negative effect. Salahodjaev et al. (2022) used a cross-country panel data to examine the relationship between environmental sustainability and renewable energy consumption. Their findings indicated that renewable energy consumption had a positive effect on environmental sustainability, while non-renewable energy consumption had a negative effect. Hieu and Mai (2023) used MMQR estimations to investigate the impact of renewable energy on economic growth in developing countries. Their results showed that renewable energy consumption had a positive effect on economic growth, while non-renewable energy consumption had a negative effect. Ping and Shah (2022) used a panel data approach to examine the relationship between green finance, renewable energy, financial development, foreign direct investment, and carbon dioxide emissions. Their results showed that renewable energy had a positive effect on financial development, foreign direct investment, and carbon dioxide emissions, while green finance had a negative effect. Kartal et al. (2023) used a nonlinear ARDL and frequency domain causality model to investigate the asymmetric effect of political stability on production-based carbon dioxide emissions in the UK. Their findings indicated that political stability had a positive effect on carbon dioxide emissions, while renewable energy consumption had a negative effect. Overall, the findings of this literature review suggest that renewable energy consumption has a positive effect on economic growth, financial development, environmental quality, and carbon dioxide emissions, while non-renewable energy consumption has a negative effect. Furthermore, the results indicate that trade openness, green finance, and political stability have varying effects on economic growth, renewable energy consumption, and carbon dioxide emissions depending on the context. Additionally, using renewable sources of energy strengthens the energy security (Radmehr et al. 2021), which is critical in times when energy market is unstable.

Methodology

Model specification and data

In this study, dynamic panel data method was used to determine the long-run relationship between renewable energy, CO₂, urbanization, and economic growth with the data sets of 12 BSEC countries. The explanatory information related to the data set and variables used for econometric analyses is presented in Table 1.

Table 1.

Definition of the variables

Variables	Definitions	Source
LRE	Natural logarithm of renewable electricity output (% of total electricity output)	International Energy Agency (IEA)
LCO2	Natural logarithm of CO2 emission (metric tons per capita)	World Bank
LUR	Natural logarithm of urban population growth (annual %)	World Bank Data
LY	Natural logarithm of GDP per capita	World Bank Data

Table 1 shows the full logarithmic model created with variables that have undergone logarithmic transformations, which is shown in Eq. (1).

$${\mathrm{LRE}}_{\mathit{it}}={\alpha }_i+{\beta }_{1i}{\mathrm{LY}}_{\mathit{it}}+{\beta }_{2i}{\mathrm{LUR}}_{\mathit{it}}+{\beta }_{3i}{\mathrm{LCO}2}_{\mathit{it}}+u_{\mathit{it}}$$ 1

(i = 1,…12) ve (t = 1995,…, 2020)

In the model, i represents the cross-sectional dimension, and t represents time dimension. Table 2 includes the 12 selected BSEC countries.

Table 2.

BSEC countries included in the analysis

N	Country	N	Country
1	Albania	7	Moldova
2	Armenia	8	Romania
3	Azerbaijan	9	Russia
4	Bulgaria	10	Serbia
5	Georgia	11	Turkey
6	Greece	12	Ukraine

Method

In this study, dynamic panel data method was used to determine the long-run relationship between renewable energy, CO₂, urbanization, and economic growth with the data sets of 12 BSEC countries. In this study, Durbin-Hausman (Durbin-H) Panel Co-integration Test, which is the second-generation panel co-integration tests, proposed by Westerlund (2008), was used to determine cointegration between variables. Before conducting the panel co-integration analysis, horizontal cross-sectional dependence and homogeneity tests must be performed. Tests proposed by Breusch and Pagan (1980), Pesaran et al. (2004), and Pesaran et al. (2008) were used to detect cross-sectional dependence. The test statistic is calculated as in Eq. (2) in the study proposed by Breusch and Pagan (1980) (Pesaran et al. 2008):

$$\begin{array}{cc}\mathrm{LM}=T{\sum }_{i=1}^{n-1}{\sum }_{j=i+1P_{\mathit{ij}}^2}^N,& \sim X^{2}N(N-1)/2\end{array}$$ 2

The LM test is valid in cases where the dimension of N is small and the dimension of T is large. The test statistic is developed by Pesaran et al. (2004) and is found in Eq. (3).

$$\mathrm{CD}=\sqrt{\frac{2T}{N(N-1)}}\left(\sum _{i=1}^{n-1},\sum _{j=i+1P_{\mathit{ij}}^2}^N,.\right)$$ 3

Under the null hypothesis, when T is sufficiently large, the limit of the CD → (0,1) function is N → ∞. In the case where T → ∞ and then N → ∞ for large panels, Pesaran et al. (2008) suggest a corrected version of the LM test. The corrected LM test is expressed as follows:

$$\begin{array}{cc}{\mathrm{LM}}_{\mathrm{adj}}=\sqrt{\frac{2T}{N(N-1)}}\left(\sum _{i=1}^{n-1},\sum _{j=i+1P_{\mathit{ij}}^2}^N,.\right)\frac{(T-k){\tilde{\rho }}_{\mathit{ij}}^2-{\mu }_{\mathit{Tij}}}{\sqrt{v_{\mathit{ij}}^2}}& \sim N(0,1)\end{array}$$ 4

Here, k is the number of regressors, and ${\mu }_{\mathit{Tij}}$ and $v_{\mathit{ij}}^2$ are the mean and variance respectively of $(T-k){\tilde{\rho }}_{\mathrm{ij}}^2$ developed by Pesaran et al. (2008).

In horizontal section dependence tests, the hypotheses are:

• H0: “There is no dependence between sections.”

• H1: “There is dependence between sections.”

According to the test results, if the H0 hypothesis cannot be rejected, the analysis continues with first-generation panel unit root tests. However, if the H0 hypothesis is rejected, it will be corrected to continue the analysis with second-generation panel unit root tests (Baltagi and Baltagi 2008).

One of the second-generation panel unit root tests, the panel unit root test proposed by Smith et al. (2004), is used to investigate the presence of unit roots in variables. This test, which takes horizontal section dependence into account, the t® test proposed by Smith et al. (2004), is the bootstrap distribution version of the panel unit root test developed by Im et al. (2003). The hypotheses of the test are:

• H0: “The series contain unit roots.”

• H1: “The series are stationary.”

Another pre-test is the delta test developed by Pesaran and Yamagata (2008) which tests whether the slope coefficients are homogeneous.

$$\tilde{\mathrm{\Delta }}=\sqrt{N}\left(\frac{N^{-1}\tilde{S}-k}{\sqrt{2k}}\right)$$ 5

The test statistic is calculated in the following way, even though the Delta test has an asymptotic normal distribution:

$$\tilde{\mathrm{\Delta }}{\mathrm{\Delta }}_{\mathrm{adj}}=\sqrt{N}\left(\frac{N^{-1}\tilde{S}-E({\tilde{Z}}_{\mathit{iT}})}{\sqrt{\mathrm{Var}({\tilde{Z}}_{\mathit{iT}})}}\right)$$ 6

In the equation above, the mean of $E({\tilde{Z}}_{\mathit{iT}})=k$, and the variance of.

$\mathrm{Var}({\tilde{Z}}_{\mathit{iT}})=\left(\frac{2k(T-k-1)}{T+1}\right)$. In the homogeneity test, the hypotheses are:

• H0: “Slope coefficients are homogeneous.”

• H1: “Slope coefficients are heterogeneous.”

The relationship between the variables is analyzed using the Durbin-Hausman (Durbin-H) Panel Co-integration Test proposed by Westerlund (2008), which can be used in the case of the dependent variable being I(1) and the independent variables being either I(0) or I(1). The null hypothesis of the test is “H0: There is no co-integration relationship,” and the alternative hypothesis is “H1: There is a co-integration relationship.” In Westerlund (2008) Durbin-H method, co-integration relationship is estimated separately for the two cases where the slope coefficient is heterogeneous (group) or homogeneous (panel) (Iorio and Fachin 2022). The long-term coefficients of the variables are estimated using the Common Correlated Effects (CCE) method proposed by Pesaran (2006), which takes into account heterogeneity and cross-sectional dependence.

Findings and discussion of the results

In this section, empirical findings are presented. The results of the cross-section dependency test results are indicated in Table 3.

Table 3.

Cross-section dependency test

Variables	LRE		LY		LUR		LCO2
Tests	Test statistic	p value	Test statistic	p value	Test statistic	p value	Test statistic	p value
CDlm1 (Breusch and Pagan 1980)	118.100	0.000	130.287	0.003	105.016	0.000	116.172	0.004
CDlm2 (Pesaran et al. 2004)	3.081	0.000	3.885	0.000	10.041	0.000	2.126	0.001
CDlm3 (Pesaran et al. 2004)	− 1.173	0.023	− 1.420	0.074	− 1.114	0.015	− 1.805	0.000
LMadj (Pesaran and Yamagata 2008)	9.717	0.000	11.261	0.000	0.288	0.234	− 0.334	0.561
Cointegration results
	Test statistic				p value
CDlm1 (Breusch and Pagan 1980)	136.546***				0.000
CDlm2 (Pesaran et al. 2004)	4.466***				0.000
CDlm3 (Pesaran et al. 2004)	5.560***				0.000
LMadj (Pesaran and Yamagata 2008)	13.861***				0.000

The findings presented in Table 3 indicate the presence of cross-sectional dependence both in variables and in the integration equation of equality at the 1% significance level. The existence of cross-sectional dependence suggests that the unit root presence in the series is estimated using the panel unit root test proposed by Smith et al. (2004). The results of the panel unit root test are presented in Table 4.

Table 4.

Smith et al. (2004) panel unit root test results

Model	Level		First difference
	Constant	Constant-trend	Constant	Constant-trend
Variables	Test statistics	Test statistics	Test statistics	Test statistics
LRE	− 0.585 (0.882)	− 1.370 (0.868)	− 5.581*** (0.002)	− 5.634*** (0.000)
LY	− 0.302 (0.878)	− 2.116 (0.421)	− 6.062*** (0.000)	− 6.041*** (0.000)
LUR	− 1.761* (0.067)	− 2.154 (0.305)	− 7.212*** (0.000)	− 7.211*** (0.000)
LCO2	− 1.052 (0.811)	− 2.011 (0.573)	− 7.531*** (0.000)	− 7.523*** (0.000)

The lag length is assumed to be 1, and the probability values are obtained from a 1000 bootstrap distribution, indicated in parentheses. The “***” symbol indicates significance at the 1% level, the “**” symbol indicates significance at the 5% level, and the “*” symbol indicates significance at the 10% level

When the results of the panel unit root test are examined, it is found that the LRE, LY, and LCO₂ variables have a unit root process at the level, but the LUR variable is stationary at the level with a significance of 10%. The results obtained require performing a second-generation panel cointegration test. However, it is necessary to determine first whether the cointegration coefficient of the model is homogeneous or heterogeneous. The results of the slope homogeneity test are in Table 5.

Table 5.

Homogeneity test results (H_o: slope coefficients are homogeneous)

Tests	Statistics	p value
Detla_tilde	12.481	0.000***
Detla_tilde_adj	13.378	0.000***

The “***” symbol indicates significance at the 1% level, the “**” symbol indicates significance at the 5% level, and the “*” symbol indicates significance at the 10% level

The results of the homogeneity tests indicate that the null hypothesis, which expresses the homogeneity of the model in the tests, is rejected at a significance level of 1%, revealing the heterogeneity of the cointegration equation. This implies that the effect of renewable energy production on indicators of economic growth, urbanization rate, and carbon dioxide emissions shows differences in every sector. The unit root test applied revealed that the LRE, LY, and LCO₂ variables contain a unit root at the level, but the LUR variable is stationary at the level, allowing the implementation of the Westerlund-Durbin-Hausman (2008) cointegration test. The results of the Westerlund-Durbin-Hausman (2008) test are given in Table 6.

Table 6.

Westerlund-Durbin-Hausman (2008) cointegration test result

Test	Statistics	p value
Durbin-Hausman Group Statistic	− 1.002	0.012**
Durbin-Hausman Panel Statistic	− 1.772	0.020**

The “***” symbol indicates significance at the 1% level, the “**” symbol indicates significance at the 5% level, and the “*” symbol indicates significance at the 10% level

According to the results of the slope homogeneity test, one of the preliminary tests performed, the detection of heterogeneous cointegration equation coefficients indicates that the group statistics should be based on the results of the Westerlund-Durbin-Hausman (2008) cointegration test. Therefore, the H0 hypothesis, which states that there is no cointegration relationship, was rejected at the 5% significance level, indicating the presence of a long-term relationship between the dependent and independent variables. After detecting a long-term relationship between the series, the cointegration coefficients are analyzed with the help of the CCE method developed by Pesaran (2006). The cointegration coefficients of the model for developing countries are given in Table 7 with the estimation results.

Table 7.

Co-integration coefficient estimation (CCE)

	LRE = f(LY)				LRE = f(LUR)			LRE = f(LCO2)
	Coef	Std. error		p value	Coef	Std. error	p value	Coef	Std. error	p value
CCE	− 0.162	0.205	0.472		0.037	0.586	0.834	− 0.123	0.533	0.748
Countries result
Albania	0.037	0.771	0.845		− 0.646	0.607	0.174	1.232*	0.702	0.082
Armenia	1.228	1.010	0.180		− 2.256***	0.881	0.002	1.042***	0.516	0.002
Azerbaijan	− 0.326***	0.144	0.004		− 1.015***	0.283	0.008	0.185	0.482	0.631
Bulgaria	− 1.362**	0.611	0.031		0.711	0.612	0.142	− 1.646***	0.523	0.000
Georgia	− 0.144	0.325	0.611		− 0.142	0.137	0.202	1.326***	0.388	0.003
Greece	− 1.067***	0.254	0.004		− 0.317*	0.110	0.057	− 0.244	0.302	0.267
Moldova	0.141	0.155	0.118		0.010	0.473	0.861	− 0.062	0.527	0.811
Romania	− 0.532***	0.114	0.003		− 1.407	0.822	0.105	1.165	1.560	0.371
Russia	− 1.651***	0.413	0.002		− 2.202**	1.161	0.038	− 0.841	0.611	0.176
Serbia	1.640	1.575	0.188		1.172*	0.552	0.076	3.474*	1.826	0.063
Turkey	− 1.432***	0.414	0.002		2.753***	1.368	0.008	− 3.856	2.131	0.114
Ukraine	0.171	0.180	0.422		− 1.557***	0.312	0.000	0.658	0.362	0.103

The “***” symbol indicates significance at the 1% level, the “**” symbol indicates significance at the 5% level, and the “*” symbol indicates significance at the 10% level

When the cointegration estimator results in Table 7 are examined, there is a negative and statistically significant relationship between renewable energy production and GDP per capita in Romania, Azerbaijan, Russia, Turkey, Bulgaria, and Greece. In these countries, a 1% increase in GDP per capita leads to a decrease in renewable energy production by approximately 0.53%, 0.33%, 1.65%, 1.43%, 1.36%, and 1.07%. The explanation may be that the renewable energy out production is not keeping the pace with the economic growth or that these national economies already reached their renewable potential. There was a statistically significant relationship between renewable energy production and urban population in Armenia, Azerbaijan, Greece, Russia, Serbia, Turkey, and Ukraine. A 1% increase in urban population leads to a decrease in renewable energy production by 2.23%, 1.02%, 0.31%, 2.20%, 1.56%, 1.17%, and 2.75% in these countries. These results describe economies where the fossil fuels are still largely the main source that keep the economic growth running.

The relationship between renewable energy production and carbon dioxide emissions was found to be significant in Albania, Armenia, Bulgaria, Georgia, and Serbia. A 1% increase in renewable energy production leads to a decrease in CO₂ emissions by 1.65% in Bulgaria, and an increase by 1.23%, 1.04%, 1.33%, and 3.47% in Albania, Armenia, Georgia, and Serbia, respectively, in line with the findings of several previous studies (Raihan et al. 2022; Grodzicki & Jankiewicz 2022).

Conclusion and recommendations

The long-run panel cointegration analysis of 12 developing countries, including Albania, Armenia, Azerbaijan, Bulgaria, Georgia, Greece, Moldova, Romania, Russia, Serbia, Turkey, and Ukraine, was performed using dynamic panel data analysis methods. The globalization process, triggered by the reduction of capital constraints to a minimal level since the 1980s, has affected the countries in many ways. It is important to measure the interdependence of these countries in empirical analyses (e.g., supply chains and foreign direct investments) to obtain robust results. Cross-sectional dependence was measured using the CDlm1, CDlm2, CDlm3, and LMadj tests. Second-generation unit root tests and cointegration tests were applied to reflect current practices. Results from the Durbin-H Co-Integration analysis (Westerlund 2008) suggest a long-run relationship between renewable energy production and GDP per capita, urban population growth rate, and carbon dioxide emissions per capita.

The results of the CCE co-integration coefficients of Pesaran (2006) showed a negative and statistically significant relationship between renewable energy production and GDP per capita in Romania, Azerbaijan, Russia, Turkey, Bulgaria, and Greece. In these countries, a 1% increase in GDP per capita leads to a decrease in renewable energy production by approximately 0.53%, 0.33%, 1.65%, 1.43%, 1.36%, and 1.07%. There was a statistically significant relationship between renewable energy production and urban population in Armenia, Azerbaijan, Greece, Russia, Serbia, Turkey, and Ukraine. A 1% increase in urban population leads to a decrease in renewable energy production by 2.23%, 1.02%, 0.31%, 2.20%, 1.56%, 1.17%, and 2.75% in these countries. These could happen due to the fact that the urbanization process is still developing in the case of some of the countries, only Greece and Russia having the share of urban population close to the region’s average. Additionally, these economies are connected to fossil fuels sources or leading producers, which tend to decrease their interest in switching to renewable sources.

The relationship between renewable energy production and carbon dioxide emissions was found to be significant in Albania, Armenia, Bulgaria, Georgia, and Serbia. A 1% increase in CO₂ emissions leads to a decrease in renewable energy production by 1.65% in Bulgaria and an increase by 1.23%, 1.04%, 1.33%, and 3.47% in Albania, Armenia, Georgia, and Serbia, respectively.

The panel results of 12 developing countries indicate that renewable energy production and the other variables in question have a long-term relationship, but the significance of renewable energy for the future is not fully understood. For example, when evaluating the results of Russia, which has become a major production center in the world and consumes a significant amount of energy due to its large trade volume and population, it has been determined that economic growth has a negative impact on renewable energy production. This is also true in Turkey and other countries included in the analysis. Despite current efforts to promote renewable energy, particularly in Russia and Turkey, the empirical findings suggest that more investments and incentives in this direction are needed. Furthermore, developed economies, which have relocated production to developing countries for economic reasons, may face the negative consequences of relying on traditional energy sources to an even greater extent. To ensure a more livable and cleaner world for future generations, the use of primary energy sources, which contribute to natural disasters, should be reduced, and serious steps should be taken to adopt renewable energy sources instead.

The energy market is anticipated to grow more fragmented and diversified, with a greater variety of sources and providers, as the globe moves toward renewable energy. This can lessen the reliance on fossil fuels and improve the energy system’s resilience to climate-related shocks like extreme weather conditions.

In addition, switching to renewable energy can lessen climate change’s effects by cutting greenhouse gas emissions. Increased usage of renewable energy can reduce CO₂ emissions and other dangerous pollutants that fuel climate change in BSEC member states.

By decreasing dependency on energy imports and boosting the accessibility of local and decentralized energy sources, the move toward renewable energy can also improve energy security. This can be crucial in light of supply chain interruptions caused by climate change.

In general, investigating the connection between urbanization, CO₂ emissions, and the generation of renewable energy in BSEC member nations has significant ramifications for energy markets; however, it is also essential for solving the urgent problem of climate change.

Based on the results of the long-run panel cointegration analysis of 12 developing countries, it is recommended that:

• More investments and incentives are needed in the promotion of renewable energy, particularly in countries like Russia, Turkey, and others where economic growth has a negative impact on renewable energy production in order to mitigate the risk in climate change.

• Developed economies that have relocated production to developing countries should be aware of the negative consequences of relying on traditional energy sources and should take serious steps toward adopting renewable energy sources.

• BSEC member states should prioritize investment in renewable energy sources as they provide a long-term solution to risk in climate change and reduce carbon dioxide emissions.

These recommendations are based on the understanding that renewable energy production, urbanization, economic growth, and carbon dioxide emissions are closely interconnected, and efforts to promote one will inevitably have an impact on the others. By working together, BSEC member states can minimize the impact of the climate change risk, reduce carbon dioxide emissions, and promote economic growth.

Based on the findings, future work could include:

• Analysis of policies and programs that have been implemented or planned to promote renewable energy production and evaluate their effectiveness.

• Investigation of the trade-offs and synergies between renewable energy production and other economic, social, and environmental objectives in each country.

The study is limited to the sample countries and variables used. Future studies may include other interesting and closely related factors such as technological innovations to establish such research.

Author contribution

Formal analysis was performed by Azer Dilanchiev, interpretation of results was done by Florian Nuta, and formal writing was done by Itbar Khan; final proof reading, writing, supervision, and conclusions were drawn by Hayat Khan.

Data availability

Data used in this paper is available upon request from the corresponding author.

Declarations

Ethical approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.